diff options
Diffstat (limited to 'src/_posts/2018-08-15-the-perl-conference-in-glasgow.adoc')
| -rw-r--r-- | src/_posts/2018-08-15-the-perl-conference-in-glasgow.adoc | 240 |
1 files changed, 240 insertions, 0 deletions
diff --git a/src/_posts/2018-08-15-the-perl-conference-in-glasgow.adoc b/src/_posts/2018-08-15-the-perl-conference-in-glasgow.adoc new file mode 100644 index 0000000..6fd8f90 --- /dev/null +++ b/src/_posts/2018-08-15-the-perl-conference-in-glasgow.adoc @@ -0,0 +1,240 @@ +--- +date: 2018-08-23 +tags: Perl Conference +description: > + My feedback and comments on some of the talks I attended during The Perl + Conference in Glasgow, in 2018. +--- += The Perl Conference in Glasgow +:toc: preamble + +This year the European Perl Conference was hosted in Glasgow, and of course +I've attended a number of presentations there. On some of these, I have some +feedback or comments. These talks, and the feedback I have for them, are +detailed in this blog post. + +== Discourse Without Drama + +[NOTE] +==== +There used to be comments here to show a different perspective to the speaker. +However, someone's lightning talk has been removed because +https://act.perlconference.org/tpc-2018-glasgow/news/1568[some people played +the "offended" card], and no specifics are being published about it. Therefore, +I cannot make any assertions on what is and what is not allowed, and have to +assume that criticism of harmful beliefs will be shut down. As such, I have +removed the comments which used to be here, in an attempt to keep it from +detracting from the rest of my post. +==== + +== European Perl Mongers Organiser's Forum 2018 + +The Perl community isn't big nowadays, however, the Perl 6 language also offers +a lot of concepts which are very well suited for modern programming. Sadly, if +no new users try out the language, it will be all for nothing. As such, we need +to bring new blood in to the community. + +One of the ways of doing this is by extending our promoting efforts outside of +the Perl community. Most people who like Perl are in a social bubble with other +people that are also familiar with the Perl programming language, be it 5 or 6. +But we need to reach new people as well, who will most likely be outside of +this social bubble. These people don't have to be techies either, they might +just as well be marketeers or designers. + +I myself am part of the "techies", so I'll stick to this particular group for +now. And I know people like me can be found at meetups, so it would be +worthwhile to promote Perl at meetups which are not dedicated to Perl. Think of +more generic programming meetups, or GNU+Linux User Groups. We have to be +mindful not to be too pushy, though. Listen to other people, and try to +understand the problem they're facing. Most of them will not be open to using a +different language immediately, especially not Perl (which sadly has a +particularly bad standing amongst people unfamiliar with it). Try to assist +them with their issues, and slowly introduce them to Perl (6) if it helps to +showcase what you mean. It might also be interesting to show people examples on +how to solve certain issues before telling them the language's name, so they +don't have a negative preconception solely from the name. + +Another thing to note is that Perl is more than just a programming language. +It's a community, and a large library of modules, known as CPAN. And CPAN +offers some nifty tools, such as the CPAN testers, which help ensure module +developers that their code runs on a massive set of platforms and Perl +versions. + +This has led me to consider the creation of a new Perl 6 module: +`CPAN::Tester`, to make it easy for people to contribute to a large-scale +testing environment for Perl 6. The idea is that one can run `CPAN::Tester` on +their machine, which will keep track of new Perl 6 modules being uploaded to +CPAN. The results are to be sent to another server (or multiple servers), which +can aggregate the data and show a matrix of test results. This aggregating +server could also be built as a Perl 6 module, possibly named +`CPAN::Tester::ResultsServer`. This would make setting up an environment +similar to CPAN testers for Perl 5 quite easy for Perl 6. + +== Perl 6 in Real Life $Work + +The speaker shows the perfect use case for +https://docs.perl6.org/language/grammars[Perl 6 grammars], advanced yet +readable parsing of text and performing actions with the results. It's an +interesting talk, showcasing some nifty grammar constructs. The best part of +this is that it actually runs in production, where it parses over 700 files, +consisting over 100,000 lines of code, in about 22 seconds (on his laptop). +This goes to show that Perl 6 is no longer "too slow to use in production". + +It might be interesting to run this application of grammars on every Perl 6 +release to gather more information on the speed improvements of Perl 6, much +like Tux's `Text::CSV` runs. + +== Releasing a Perl 6 Module + +The speaker starts off with detailing the platform which most Perl 6 modules +use to host their code repository, GitHub. He also touched upon automated +testing using Travis and AppVeyor. It was good to show how to make use of +these, as automated testing oftentimes stops unintended bugs from reaching end +users. But, I personally prefer GitLab over GitHub, as they have much better +testing functionality, and they actually release their own platform as an open +source package. I'd like more GitLab love from the community and speakers as +well if possible. This would also make the speaker's CI configuration simpler, +for which he currently uses a `.travis.yml` file. This requires him to build +Perl 6 from source every test run, wasting quite a lot of time. + +It was also noted that there's a module to help you set up this module +skeleton, `mi6`. The speaker also noted that it doesn't seem to add much once +you know how a Perl 6 module is organized, and I tend to agree with this. +Actually, I made a module precisely because I agree with him here, +`App::Assixt`. This module intends to smoothen the entire course of module +development, not just the creation of a skeleton file. It will take care of +keeping your `META6.json` up to date, and ease uploading your module to CPAN as +well. + +Lastly, the speaker says the `META6.json` documentation can be found in S22. +While this is technically correct, S22 is *not* the implementation's +documentation, this lives in the official Perl 6 documentation instead. S22 +offers many additional information to be stored in the `META6.json`, but using +these fields will actually break installation of your module through `zef`, +rendering it unusable by others. I would strongly recommend people not to use +S22 when trying to figure out what they can or cannot do with their +`META6.json`. + +== How to become CPAN contributor? + +Submitting a pull request (or more correctly named, merge request) to a +repository is possibly the most straightforward way to help out other projects. +However, sometimes it will take a long time to get a response. The speaker +notes this can actually be on the scale of years. I have authored a number of +modules myself, and have been in the situation where I had not realized I got a +merge request from another person (same goes for issue reports). I would +recommend people who are not getting timely responses to their contributions to +contact the maintainer via other channels which are more suited for +communications. Think of email or IRC, for instance. You'll generally have a +much better chance of getting a timely response from the author, and then you +can work out your contribution and see if you can get it merged into the main +project. + +The speaker also lists a couple of ways to get started with contributing to +modules. One thing I missed in particular was the Squashathons +footnote:[A Squashathon is like a hackathon, except everyone in the world is +invited, and you can help out over the Internet, staying in your own home. Of +course, you can still meet up with other developers and make it a social +gathering in the real world as well!] for Perl 6. +These generally offer a good entry point to help out with the language's +development and the ecosystem's maintainance. + +Near the end, it was pointed out that it is a good idea to have a thick skin. +Even when it's not intended, people can come accross as rude. This is in +opposition to the talking point of the speaker yesterday (_Discourse Without +Drama_), but he does raise a good point here. People oftentimes don't mean to +insult you, but context is easily lost in written communications. Try to stay +mature and professional, you can simply ask for clarification. If you feel the +person remains hostile towards you, walk away. There's plenty of other projects +that would love your contributions! + +== Conference Organizers & European Perl Mongers Organiser's Forum 2018 BoF + +Well, that's certainly a mouthful for a heading, and it even contains an +abbreviation! This event was not a presentation, but a platform to exchange +ideas together. + +One of the items that were up for discussion was _A Conference Toolkit_, or ACT +for short. This is the platform used to organize Perl events, such as this +conference and Perl workshops throughout the world. However, ACT is dated. +They enabled HTTPS a short while ago, but it's still not the default because +people don't want to risk breaking the platform. I think this is enough of +an indication that it might be time to make something new to replace it. + +And I'm not alone in that sentiment, it seems. However, ACT is big and contains +a lot of data we don't want to lose. It's a massive undertaking to make a new +tool that works at least as well, and allows us to make use of the old data as +well. There is a Trello board available that lists all the features that would +be required to implement, so that's a good start already. I think now it needs +a dedicated product owner with people contributing code, so a start can be +made. This does seem like a touchy subject, since I'm far from the first person +to want this. Many before me have tried and failed already. + +As such, I'd propose not making it a Perl centric tool. Make it a modular, +generic event organizing tool. Get a good database design that we can import +our old data into, so nothing is lost, but things can be converted to be more +useful for our current needs. This way, we can work in small steps, and maybe +even reach contributors from outside the regular Perl circles. This might even +bring in new partnerships (or sponsors) towards the Perl community. + +Personally, I'd like to see something like this to be written in Perl 6. This +way, it could also be used as a showcase project for the Perl 6 programming +language. + +== Writing a Perl 6 Module + +Perl 6 has this very neat feature called +https://docs.perl6.org/language/typesystem#index-entry-subset-subset[subsets]. +These can be used to make your own types with very little effort, which can +help tremendously to keep your code clean and concise. There are two arguments +I have in favour of subsets that the speaker did not touch upon. + +First off, using a subset instead of a `where` clause in a sub or method +signature will bring much better error messages. If you use a `where` in your +signature, and the check fails, you'll get an error that there was no signature +that matched `where { ... }`. + +Secondly, if you want to use abstract methods, you can't really use a `where`. +https://stackoverflow.com/questions/51570655/how-to-use-abstract-multi-methods-containing-a-where[I've +asked a question about this on Stack Overflow], which has the details as to why +this doesn't work the way you might expect. + +Next, there's some cool things about operators in Perl 6. There are many of +these available by default, and it's _very_ easy to add new ones yourself as +well. In fact, the `Math::Matrix` module used throughout the presentation makes +some available as well. Thanks to the ease of adding operators in Perl 6, if +you have a `Math::Matrix $m` in Perl 6, you can get the norm by writing `|| $m +||`. This is the mathematically correct way to write this, making it easy to +understand for everyone using matrixes in their daily lives. If you're a +mathematician, small things like these are great to have. + +I have some comments on the `Math::Matrix` module itself as well, based on +slides shown in the presentiation. The first thing I noticed is that there's a +`norm` method using a `where` clause when it's not needed: + +[source,perl6] +---- +method norm (Str $which where * eq 'row-sum') +---- + +This can be written instead as: + +[source,perl6] +---- +method norm ('row-sum') +---- + +This is shorter and clearer, and you'll get better feedback from the compiler +as well. I https://github.com/pierre-vigier/Perl6-Math-Matrix/pull/49[submitted +a pull request on the GitHub repository] in an attempt to improve this, which +got merged! The speaker was not aware it could be done in this manner, so I'm +proud I got to teach him something right after he did his presentation. + +== Winding down + +I've had a great time at the Perl conference, spoke to many people with whom +I've had some great discussions. I got to meet and personally thank a number of +people who've helped me out over the past year as well. + +A big thank you to all the people who made this conference possible, and I hope +to see you all again in Riga! |